Two Subwoofers... Comb Effect

Duke, your room sounded great at RMAF; I especially loved the clean off-axis presentation! Kudos to you for thinking outside the (subwoofer) box!

I should clarify that the punchy feeling of coherent bass is a kind of compression - it is not something so much as you hear in the sense of a note but more like something you feel. I find if I place the subwoofer too far away from the mains (for example the middle of the room or behind the listener then the notes are still audible but the "attack" is less...I suspect the transient response just becomes wrong...the hit from the beater on the drum head arrives later than the bass by however many feet you are closer to the sub than the speaker (6 feet = 6 millisecs and I think we can begin to perceive this even with bass frequencies.)

Hope this clarifies - I certainly agree that for the sound of the note itself it does not seem to matter where it is placed - so getting an even response makes sense.

If you have a DSP then you can set the distance (sub to listener) and have the processor delay the output to the sub to maintain phase. However this may not work for several listeners seated at different distances to sub and speakers...

HSU Research recommends 'near field' sub placement. When I was up there (OC, SoCal) I sat in there listening/audition area and sat virtually next to the sub(s) under test. Still difficult to localize and the 'blend' was good.
If I could work it out, I'd TRY my sub in back of the sofa, concealed (sort of) under the sofa back table. However, I got lucky and near corner placement (3 feet out) and against a long wall made it happen.
I have the advantage, by sheer accident, of having a room with 8 or so walls, 3 45degree corners and an off-center asymmetric ceiling of 12' max and 8' at one side. This room seems to break up or not be conducive to the formation of 'bad' standing waves....(so called peaks or suckouts).

Treatments ARE in the offing, but the idea would be to kill first reflection points from my Maggies and deaden the back of the room....the back wall is a reflection monster, too.

MAAB, please be careful, and Shad, this is meant with all due respect, of any and MOST prescriptive advice. Your listening area is a unique problem and Opportunity.
Experiment around. Even if you can't leave something there due to WAF, try it. Send everyone OUT for the day. Invest in some movie tickets for 'em all. Get RID of them so you can move stuff around and experiment. That's how I discovered, many years ago, that the one hard/fast rule of Magnepans.....IE, tweeters OUT doesn't apply in ALL instances. You will learn stuff nobody could tell you about your room / system / ears / interactions.
It'll all be good.

I think it's an advantage to place the subs as close to the mains as practical running in mono to minimize phase issues. There is going to be some overlap of frequency and you don't want the same frequency coming from sources several feet away from each other. Speaker makers go to alot of trouble to align drivers. Depending on the crossover slope it might also be beneficial to angle them within 15 degrees or so of the mains. Yes it's all room dependent and you could end up with something quite different but I would start with the sub drivers and the main drivers on the same plane as close together as possible with a similar angle. In a normal sized room treatments are likely to be needed - Jim

A reason to place stereo subs as close to mains as possible is the inevitable "leakage" of sound above the crossover level. On paper, a driver contributes 50% to overall spl at crossover frequency and a sub continues to "roll off" above from there.
Listen to your system with only the subs powered up, don't even need an spl meter.
All that "supra-sub" content from the subs would seem to affect imaging, etc. the least by such positioning, though I've not done subjective testing. My stereo subs will stay where they are - immediately adjacent to each main speaker.
A Velodyne SMS-1, Rives PARC and other of their ilk will do a reasonable job smoothing peaks to make the nulls less toublesome at one's listening position.